curl_url_set with CURLUPART_URL with URL already set rejects empty string

[zopsicle]

I did this

auto url = curl_url();
curl_url_set(url, CURLUPART_URL, "https://example.com", 0);
auto result = curl_url_set(url, CURLUPART_URL, "", 0);
assert(result == CURLUE_OK);

I expected the following

As per the documentation for curl_url_set:

If the handle already is populated with a URL, the new URL can be relative to the previous.

An empty string should be considered a relative URL which, when made absolute given another URL, leaves that URL untouched.

Instead, curl_url_set returns CURLUE_MALFORMED_INPUT.

curl/libcurl version

curl 8.12.1 (aarch64-unknown-linux-gnu) libcurl/8.12.1 OpenSSL/3.3.3 zlib/1.3.1 brotli/1.1.0 zstd/1.5.6 libidn2/2.3.7 libpsl/0.21.5 libssh2/1.11.1 nghttp2/1.64.0
Release-Date: 2025-02-13
Protocols: dict file ftp ftps gopher gophers http https imap imaps ipfs ipns mqtt pop3 pop3s rtsp scp sftp smb smbs smtp smtps telnet tftp
Features: alt-svc AsynchDNS brotli GSS-API HSTS HTTP2 HTTPS-proxy IDN IPv6 Kerberos Largefile libz NTLM PSL SPNEGO SSL threadsafe TLS-SRP UnixSockets zstd

operating system

Linux utm 6.6.85 #1-NixOS SMP Fri Mar 28 20:59:57 UTC 2025 aarch64 GNU/Linux

[bagder]

An empty string should be considered a relative URL

Why should it be considered that?

[zopsicle]

RFC 3986 4.2. Relative Reference considers an empty string a relative reference:

      relative-ref  = relative-part [ "?" query ] [ "#" fragment ]

      relative-part = "//" authority path-abempty
                    / path-absolute
                    / path-noscheme
                    / path-empty

and the resolution algorithm (section 5.2) handles it thus:

if (R.path == "") then
    T.path = Base.path;
    ...
endif

This is also in line with the WHATWG URL spec and the behavior of web browsers when encountering <a href="".

[bagder]

Strong arguments, I agree.